Microwave-Assisted Freeze-Drying of Monoclonal Antibodies: Product Quality Aspects and Storage Stability

pharmaceutics
Article
Microwave-Assisted Freeze-Drying of Monoclonal
Antibodies: Product Quality Aspects and
Storage Stability
Julian Hendryk Gitter 1, *, Raimund Geidobler 2 , Ingo Presser 2 and Gerhard Winter 1, *
1 Department of Pharmacy, Pharmaceutical Technology and Biopharmaceutics,
Ludwig-Maximilians-Universität München, 81377 Munich, Germany
2 Boehringer Ingelheim Pharma GmbH & Co. KG, Pharmaceutical Development Biologicals, 88397 Biberach
an der Riß, Germany; [email protected] (R.G.);
[email protected] (I.P.)
* Correspondence: [email protected] (J.H.G.); [email protected] (G.W.)

Received: 30 October 2019; Accepted: 7 December 2019; Published: 12 December 2019
Abstract: In order to overcome the downside of long conventional freeze-drying (CFD) process times
for monoclonal antibody formulations, microwave-assisted freeze-drying (MFD) was introduced.
Recently, the general applicability and potential shortening of drying times were shown. However,
little is known about the storage stability of MFD products compared to CFD references. Additionally,
batch homogeneity issues were seen within MFD in the past. In this study, we examined four different
formulations of two different monoclonal antibodies using three different glass-forming excipients:
sucrose, trehalose, and arginine phosphate. These formulations were freeze-dried with two different
drying protocols (CFD and MFD), stored for 24 weeks, and analyzed for solid-state and protein-related
quality attributes. Moreover, a new microwave generator setup was investigated for its potential
to improve batch homogeneity. In all investigated formulations, comparable stability profiles were
found, although the classical magnetron generator led to inferior batch homogeneity with respect to
residual moisture distribution. In contrast, the new MFD setup indicated the potential to approximate
batch homogeneity to the level of CFD. However, for future applications, there is an unabated need
for new machine designs to comply with pharmaceutical manufacturing requirements.
Keywords: freeze-drying; lyophilization; drying; microwave; protein; monoclonal antibody; stability
1. Introduction
Conventional freeze-drying (CFD), also referred to as lyophilization, is a gentle drying method
to improve the long-term stability of pharmaceuticals, specifically of protein drugs [1]. The method
has been used for pharmaceutical industrial purposes since World War II, for the preparation of
human blood plasma [2], and the demand for freeze-drying (FD) remains high. By 2018, one-third of
all parenteral protein formulations approved by the European Medicines Agency were freeze-dried
products [3]. During lyophilization, the protein drug is immobilized in the solid-state, slowing down
chemical and physical degradation reactions [2,4–7]. Additionally, freeze-dried solids may have other
benefits with respect to shipping and storage [8].
In general, freeze-drying comprises three steps: freezing, primary drying (= sublimation drying),
and secondary drying (= desorption drying). Typically, the sublimation step is widely described
to be the most time-consuming, and conventional freeze-drying is associated with lengthy process
times [2,9–12]. One alternative drying method utilizing microwaves is known from the food industry:
microwave-assisted freeze-drying (MFD) [13]. Here, it is specifically used for high-value goods like
dry fruit [14]. Similar to the conventional freeze-drying process, the material to be dried first needs to
Pharmaceutics 2019, 11, 674; doi:10.3390/pharmaceutics11120674 www.mdpi.com/journal/pharmaceutics

Pharmaceutics 2019, 11, 674 2 of 21
be frozen. In a second step, the drying itself takes place. In contrast to CFD, the main heat transfer
mechanism is radiation rather than convection and conduction. Especially polar substances, e.g., water,
sugars, and amino acids, show good absorption of electromagnetic waves of wavelengths of 12.2 cm
and frequencies of 2.45 GHz [15,16]. In brief, the heating mechanism in pharmaceutics occurs due to
dipolar and ionic mechanisms. When such a polar compound is placed in an oscillating field, dipoles or
ions try to realign in the direction of the electric field. Due to the ultra-rapid change in the direction of
the electric field, internal friction of the molecules is caused, leading to heating within the material, i.e.,
volumetric heating. In the case of ions, a charge-driven migration is discussed [16–18]. MFD has clear
advantages over conventional drying processes, like significantly shorter process times [19,20], and
in the field of food processing, in the maintenance of shape, color, taste, odor, and texture [14,21–23].
In the transition area between food and pharmaceutical technology, MFD was used for the gentle
drying of bacteria suspensions. Ambros et al. [19] investigated the survival rate and viability of
different bacterial cultures. They found comparable survival rates of the investigated cultures produced
by MFD compared to conventional freeze-drying but were able to shorten process times by up to
80%. The first usage in pharmaceutical freeze-drying was presented by Evans et al. [24] at the CPPR
Freeze Drying of Pharmaceuticals & Biologics Conference in 2014, showing the general applicability to
monoclonal antibodies and vaccine formulations. On this basis, a handful of international patents
were filed claiming engineering- [25] or formulation-/process-focused [26,27] intellectual property. In a
previously published work from our group [20], the general applicability to various pharmaceutical
freeze-drying excipient systems containing a monoclonal antibody was underlined. Moreover, the
potential for process drying time reductions was discussed. However, two major questions that
have been raised have not been answered yet: (1) How do different microwave-assisted freeze-dried
antibody formulations perform in accelerated stability studies with respect to solid-state and protein
stability compared to a conventionally freeze-dried reference? (2) Is the inferior batch homogeneity
found for MFD samples a general issue associated with microwave drying, or are there ways to
improve it?
The current study examines four different formulations of two different monoclonal antibodies in
the presence of three different glass-forming excipients: sucrose, trehalose, and arginine phosphate.
These formulations were freeze-dried with two different drying protocols, i.e., using conventional
freeze-drying and microwave-assisted freeze-drying. Moreover, a new microwave setup equipped
with a semiconductor solid-state microwave generator was used for one of the formulations. Samples
were stored for 24 weeks at different temperatures (2–8 ◦ C and 40 ◦ C) and analyzed at fixed times
for their solid-state and protein-related quality attributes. We hypothesize that, on the one hand,
irrespective of the monoclonal antibody formulation, comparable stability profiles can be found for
CFD and MFD. On the other hand, we anticipate the new microwave machinery setup to have a
positive effect on batch homogeneity in microwave-assisted freeze-dried products.
2. Materials and Methods
2.1. Materials
Two different IgG type 1 monoclonal antibodies (mAb) were investigated. mAb1 was kindly
provided by Boehringer Ingelheim Pharma GmbH & Co. KG (Ingelheim am Rhein, Germany). mAb2
was an on stock at Ludwig-Maximilians-Universität München (LMU).
For mAb1-formulations, the following excipients were used: ACS certified D(+) Sucrose, which
was purchased from Sigma-Aldrich (Steinheim, Germany), D(+) Trehalose dihydrate (min. 99%
purity) was obtained from VWR International BVBA (Leuven, Belgium). EMPROVE® exp L-Arginine
(Ph. Eur. certified), EMSURE® ortho-Phosphoric acid 85% and Ph. Eur. certified Tween 80® were
obtained from Merck KGaA (Darmstadt, Germany). For mAb2-formulation, EMPROVE® exp sucrose
(Ph.Eur.-certified) purchased from Merck KGaA (Darmstadt, Germany) was used.
L-Histidine monohydrochloride monohydrate (min. 99% purity) and L-Histidine (Cell culture
reagent) were purchased from Alfa Aesar (Karlsruhe, Germany). Di-sodium hydrogen phosphate
dihydrate and sodium dihydrogen phosphate dihydrate were obtained from AppliChem (Darmstadt,
Germany). Trizma® base BioXtra (>99.9%) and Trizma®hydrochloride BioXtra (>99.0%) were
purchased from Sigma-Aldrich (St. Louis, MO, USA). Sodium chloride was obtained from Bernd Kraft
(Duisburg, Germany). Sodium hydroxide was purchased from Merck KGaA (Darmstadt, Germany).
For the preparation of buffers and stock solutions, water for injection (WFI; Purelab Plus, USF
Elga, Celle, Germany) was used.
2.2. Study Design

The four different formulations, F1–F4, which were dried either by microwave-assisted
freeze-drying (MFD) or by conventional freeze-drying (CFD), were stored for 24 weeks at refrigerator
temperature 2–8 ◦ C, at 25 ◦ C (F1), and at 40 ◦ C (Table 1). The low concentration mAb formulations
(F1–F3) were produced using a previously described MFD setup with a 2 kW/2450 MHz magnetron [20],
whereas the high concentration mAb formulation (F4) was processed using a novel semiconductor
solid-state microwave radiation source tunable from 5 W to 450 W/2450 MHz.
Table 1. Formulations used in this study.
Ingredient F1 F2 F3 F4
mAb1 [g/L] 5 5 5 /
mAb2 [g/L] / / / 50
Sucrose [% (w/v] 10 / / 5
Trehalose [% (w/v)] / 10 / /
Arginine phosphate [%
/ / 10 /
(w/v)]
Polysorbate 80 [% (w/v)] 0.02 0.02 0.02 /
Formulations F1 and F2 were formulated in 10 mM histidine buffer (pH 6.0), whereas F3 contained no additional
buffer salt but was formulated and adjusted to pH 6.0. F4 was formulated in 10 mM histidine buffer (pH 5.5).
mAB = monoclonal antibody.
2.3. Preparation of Formulations

mAb1 was dialyzed against 10 mM histidine buffer (F1, F2) or 10 mM arginine phosphate (F3) at
pH 6.0 for 24 h using dialysis membranes Spectra/Por® (MWCO 6000–8000 Da; Spectrum Laboratories
Inc., Compton, CA, USA) with two buffer exchanges. After dialysis, the concentration of mAb1 was
measured with a NanoDrop™ 2000 UV photometer (Thermo Scientific, Wilmington, Delaware) at
280 nm using an extinction coefficient of ε0.1% = 1.49 g/100 mL−1 cm−1 .
mAb2 (F4) was dialyzed and concentrated using a cross-flow filtration unit Minimate™ TFF
capsule with omega polyethersulfone (PES) membrane (MWCO 30,000 Da; Pall Corporation, New York,
NY, USA) by adding a 10-fold excess of 10 mM histidine buffer (pH 5.5). After reaching the desired
volume, the concentration of the mAb was measured with a NanoDrop™ 2000 UV photometer at 280 nm
using an extinction coefficient of ε = 225,000 M−1 cm−1 and a molecular weight of MW = 145.5 kDa.
Formulations were prepared according to the composition shown in Table 1.
F1–F3 were filtered using 0.2 µm PES membrane syringe filters (VWR International, Radnor,
PA, USA), whereas F4 was filtered using a 0.22 µm PES Sartolab® RF vacuum filter unit (Sartorius
AG, Goettingen, Germany). For each formulation, 2.3 mL was filled in 10R tubing vials (MGlas AG,
Muennerstadt, Germany) and semi-stoppered with lyophilization stoppers (FluroTec® rubber stopper,
West Pharmaceuticals, Eschweiler, Germany). The vial population for conventional freeze-drying
was arranged on a lyophilization tray, and surrounded with at least one row of 10% (w/v) sucrose
shielding vials.
2.4. Freeze-Drying Process

All samples of a corresponding formulation were frozen in the same freezing step. The formulations
F1–F3 were frozen in a Christ ε2-6D laboratory-scale freeze-dryer (Martin Christ, Osterode am Harz,
Germany) with equilibration at −5 ◦ C for 1 h, followed by ramping down the shelf with 1 K/min to a
−60 ◦ C set point.
Formulation F4 was frozen in an FTS Systems LyoStar™ 3 freeze-dryer (SP Scientific, SP Scientific,
Stone Ridge, NY, USA) with equilibration at 5 ◦ C for 1 h, followed by ramping down the shelf with
1 K/min to a −50 ◦ C set point.
The frozen samples were subjected to one of the following drying protocols:
2.4.1. Conventional Freeze-Drying (CFD)

The conventional freeze-drying cycles are summarized in Table 2. The freeze-dryer used for
cycle F1 and F2/F3 was not equipped with process analytical technologies for sublimation endpoint
determination like comparative pressure measurement. Several different formulations were dried in
one run at the same time. The holding time for primary drying, therefore, was chosen to allow for
completed sublimation.
Table 2. Overview of the conventional freeze-drying (CFD) processes for the respective formulations.
Freeze-Drying Process Setpoint Freezing Primary Drying Secondary Drying

CFD cycle F1 TShelf [◦ C] 20 −5 −60 −20 0 20
Ramp [K/min] − 1 1 0.2 0.05 0.2
Hold time [min] 5 60 2580 a 2700 (1120) b − 360
pChamber [µbar] − − − 100 50 50
CFD cycle F2/3 TShelf [◦ C] 20 −5 −60 −25 0 20
Ramp [K/min] − 1 1 0.2 0.05 0.2
Hold time [min] 5 60 3610 a 5760 (1400) b - 360
pChamber [mbar] − - − 100 50 50
CFD cycle F4 TShelf [◦ C] 20 5 −50 −20 5 35
Ramp [K/min] − 1 1 1 0.15 0.3
Hold time [min] 5 60 1006 a 1868 (1058) b − 420
pChamber [mbar] − − − 100 100 100
a Freezing times were held longer than usual due to logistical reasons caused by the need to split up the batch before
proceeding with the drying process. b Estimated time needed for complete sublimation based on the time the last
thermocouple in the respective formulation needed to reach the shelf temperature setpoint.
2.4.2. Microwave-Assisted Freeze-Drying (MFD)

Microwave-assisted freeze-drying was conducted on a modified laboratory-scale Püschner
µWaveVac 0250fd vacuum dryer prototype (Püschner GmbH + Co KG, Schwanewede, Germany).
For the formulations F1–F3, the setup described previously [20] was used. Briefly, it contained
a 2 kW/2450 MHz magnetron, a condenser (−80 ◦ C), and a vacuum system comprising a root pump
and a rotary vane pump. The tuner, which was located between the magnetron and water load, was
adjusted so that approximately 1/10 of the generated microwaves went into the product chamber. For
a schematic overview of the general setup, the reader is referred to reference [28]. In the setup used for
F1–F3, water load and product cavity were interchanged. Drying was carried out at a pressure of 8 µbar
to 30 µbar measured by Pirani gauge, and at a microwave power between 23 W to 110 W measured by
a HOMER™ impedance analyzer (S-TEAM Lab, Bratislava, Slovak Republic), until a constant mass
was reached. The drying process used for F1 and F2/F3 is presented in Figure 1a,b, respectively.
Pharmaceutics 2020, 12, x FOR PEER REVIEW 5 of 21
Figure
Figure 1. Graphical overview
1. Graphical overview ofof the
the microwave-assisted
microwave-assisted freeze-drying
freeze-drying (MFD)
(MFD) processes
processes for
for (a)
(a) F1,
F1, (b)
(b)
F2/F3, and (c) F4. Microwave power input (MW Power) is the actual measured radiated
F2/F3, and (c) F4. Microwave power input (MW Power) is the actual measured radiated microwave microwave
power,
power, the
the chamber
chamber pressure
pressure isisthe
thePirani
Piranigauge
gaugereadout
readout(p
(pChamber ), and Tc represents the readout of
Chamber), and Tc represents the readout of
the glass fiber temperature probe.
the glass fiber temperature probe.
In contrast, F4 was processed with a partially modified setup comprising a semiconductor

In contrast, F4 was processed with a partially modified setup comprising a semiconductor solid-
solid-state 500 W/2450 MHz microwave radiation source tunable from 5 W to 450 W, which directly
state 500 W/2450 MHz microwave radiation source tunable from 5 W to 450 W, which directly emitted
emitted its radiation into the product chamber [28]. Moreover, the vacuum system was complemented
its radiation into the product chamber [28]. Moreover, the vacuum system was complemented by the
by the addition of a turbopump to allow for lower chamber pressures. Drying, as it is shown in
addition of a turbopump to allow for lower chamber pressures. Drying, as it is shown in Figure 1c,
Figure 1c, was carried out at a pressure of 5 µbar to 20 µbar measured by Pirani gauge and at a
was carried out at a pressure of 5 µbar to 20 µbar measured by Pirani gauge and at a microwave
microwave power between 18 W to 99 W measured by a HOMER™ impedance analyzer (S-TEAM
power between 18 W to 99 W measured by a HOMER™ impedance analyzer (S-TEAM Lab,
Lab, Bratislava, Slovak Republic).
Bratislava, Slovak Republic).
For process monitoring, two glass fiber temperature measurement probes (TS2, Weidmann
For process monitoring, two glass fiber temperature measurement probes (TS2, Weidmann
Technologies Deutschland GmbH, Dresden, Germany) were used. Stoppering of the samples was
Technologies Deutschland GmbH, Dresden, Germany) were used. Stoppering of the samples was
carried out externally in a glove bag flushed with dry nitrogen. The dried crimped samples were kept
carried out externally in a glove bag flushed with dry nitrogen. The dried crimped samples were kept
refrigerated until analysis.
refrigerated until analysis.
2.5. Karl Fischer Titration
2.5. Karl Fischer Titration
Karl Fischer titration, equipped with a headspace module, was used to determine residual water
Karl Fischer titration, equipped with a headspace module, was used to determine residual water
content after freeze-drying. Between samples, aliquots of 9 mg and 28 mg were prepared in a glove
content after freeze-drying. Between samples, aliquots of 9 mg and 28 mg were prepared in a glove
box filled with pressurized air with a relative humidity of less than 10%, placed into 2R vials, and
box filled with pressurized air with a relative humidity of less than 10%, placed into 2R vials, and
stoppered. The samples were then placed in an oven with 100 ◦ C to enable the fast extraction of
stoppered. The samples were then placed in an oven with 100 °C to enable the fast extraction of water.
water. The headspace moisture is transported into a coulometric Karl Fischer titrator (Aqua 40.00,
The headspace moisture is transported into a coulometric Karl Fischer titrator (Aqua 40.00,
Elektrochemie Halle, Halle (Saale), Germany). Results were calculated in relative water content (w/w).
Elektrochemie Halle, Halle (Saale), Germany). Results were calculated in relative water content (w/w).
For verification of equipment performance, three aliquots of Apura® water standard oven 1% by
For verification of equipment performance, three aliquots of Apura® water standard oven 1% by
Merck KGaA (Darmstadt, Germany) were measured within a sequence.
Merck KGaA (Darmstadt, Germany) were measured within a sequence.
2.6. Brunauer-Emmet-Teller Krypton Gas Adsorption

2.6. Brunauer-Emmet-Teller Krypton Gas Adsorption

The specific surface area of dried samples was determined using Brunauer–Emmet–Teller (BET)
krypton gas adsorption in a liquid nitrogen bath at 77.3 K (Autosorb 1; Quantachrome, Odelzhausen,
Germany). Samples of 90–200 mg were gently crushed with a spatula and weighed into glass tubes.
Prior to measurement, an outgassing step was performed for at least 6 h at room temperature. An
11-point gas adsorption curve was measured, covering a p/p0 ratio of approximately 0.05–0.30. Data
evaluation was performed according to the multi-point BET method fit of the Autosorb 1 software.
2.7. X-Ray Powder Diffraction

To determine the solid-state of the lyophilizates, an XRD 3000 TT diffractometer (Seifert,
Ahrensburg, Germany) was used. The device was equipped with a copper anode (40 kV, 30 mA) and
had a wavelength of 0.154178 nm. The scintillation detector voltage was 1000 V. The samples were
placed on the copper sample holder and analyzed in the range of 5–45◦ 2-theta, with steps of 0.05◦ .
2.8. Reconstitution of Lyophilizates

The lyophilized cakes were reconstituted by the addition of WFI. The WFI volume for each
formulation was calculated to match the volume of the water removed during freeze-drying.
Reconstitution time was determined by recording the time between adding the respective
formulation-specific volume of water for injection and obtaining a clear solution without visible
matter. This observation was performed by manual visual inspection. Reconstitution was performed
applying gentle swirling for 5 s directly after the addition of water.
2.9. High-Performance Size Exclusion Chromatography (HP-SEC)

A Waters 2695 Separation module (Waters GmbH, Eschborn, Germany) equipped with a Waters
2487 Dual λ Absorbance Detector (Waters GmbH, Eschborn, Germany) at 214 and 280 nm was used.
Isocratic elution with a 25 mM sodium phosphate running buffer containing 200 mM sodium chloride
(pH 7.0) was performed.
For mAb1-formulations (F1–F3), 10 µL of a reconstituted solution corresponding to a loading
of 50 µg protein were loaded on a Tosoh TSKgel G3000SWxl, 7.8 × 300 mm, 5 µm column (Tosoh
Bioscience, Griesheim, Germany) and separated with a flow rate of 0.7 mL/min.
For mAb2 (F4), samples were diluted with 10 mM histidine buffer (pH 5.5) to 1 g/L protein
concentration, and 25 µL was injected, corresponding to a load of 25 µg protein. A YMC-Pack Diol-300,
300 × 8.0 mm, 5 µm column (YMC Europe GmbH, Dinslaken, Germany) with a flow rate of 0.8 mL/min
was used for separation. Samples were measured in triplicates with three individual injections. Data
integration of relative areas at 280 nm was performed using Chromeleon 6.80 (Thermo Scientific,
Wilmington, DE, USA), provided that every peak eluting before the monomer corresponded to high
molecular weight (HMW) species. No peaks could be detected after the monomer. For verification
of equipment performance, an internal standard of thawed mAb formulation was injected at the
beginning and end of a sequence.
2.10. High-Performance Cation Exchange Chromatography (HP-CEX)

A Waters 2695 Separation module (Waters GmbH, Eschborn, Germany) equipped with a Waters
2487 Dual λ Absorbance Detector (Waters GmbH, Eschborn, Germany) at 214 and 280 nm was used for
weak cation exchange chromatography. A linear sodium chloride gradient of 0% to 20% solvent B in
solvent A over 30 min was used for elution at a flow rate of 1 mL/min. For all cation exchange (CEX)
analysis, a ProPac™ WCX-10G BioLC™ Analytical column 4 × 250 mm equipped with a ProPac™
WCX-10G BioLC™ guard column 4 × 50 mm (ThermoFisher Scientific, Waltham, MA, USA) was used.
For mAb1-formulations (F1–F3), the solvents were composed of A: 20 mM TRIS (pH 7.1) and
B: 20 mM TRIS (pH 7.1) plus 300 mM sodium chloride. Reconstituted sample aliquots of 10 µL,
corresponding to a loading of 50 µg protein, were loaded on the column.
For mAb2 (F4), the solvents were composed of A: 20 mM TRIS (pH 7.5) and B: 20 mM TRIS (pH
7.5) plus 300 mM sodium chloride. Before analysis, samples were diluted with solvent A to 1 g/L
protein concentration, and 50 µL was injected, corresponding to a load of 50 µg protein.
Samples were measured as triplicates with two individual injections. Data integration of relative
areas was performed using the Chromeleon 6.80 software (Thermo Scientific, Wilmington, DE, USA),
provided that every peak eluting before the main peak corresponded to acidic species and peaks
eluting after the main peak corresponded to basic species. For verification of equipment performance,
an internal standard of thawed mAb formulation was injected at the beginning and end of a sequence.
2.11. Light Obscuration

One method used to determine subvisible particles of the formulation F1 was light obscuration.
Therefore, a PAMAS SVSS-35 particle counter (PAMAS—Partikelmess- und Analysesysteme GmbH,
Rutesheim, Germany) equipped with an HCB-LD 25/25 sensor, which had a detection limit of
approximately 120,000 particles ≥ 1 µm per mL, was used. The pre-rinsing volume was 0.4 mL and was
followed by three measurements of 0.2 mL. The fill rate, emptying rate, and rinse rate of the syringe
were set to 10 mL/min. Before and between samples, the system was rinsed with WFI until less than 30
particles/mL ≥ 1 µm and no particles larger than 10 µm were present. Data collection was done using
PAMAS PMA software, and particle diameters in the range of ≥1 µm to 200 µm were determined. All
results are given in cumulative particles per milliliter.
2.12. Flow-Imaging Microscopy

Due to the high transparency of protein particles, an orthogonal method for subvisible particle
determination was introduced for formulations F2–F4. Flow-imaging microscopy was performed
on a FlowCAM® 8100 (Fluid Imaging Technologies, Inc., Scarborough, ME, USA) equipped with a
10× magnification cell (81 µm × 700 µm). Prior to a measurement set, the cell was cleaned with a 1%
Hellmanex III solution and WFI. For adjustment of the focus, the default autofocus procedure using
20 µm calibration beads was performed. Sample solution volumes of 150 µL were measured with a
flow rate of 0.10 mL/min, at an image rate of 29 frames per second, and an estimated run time of 1.5 min.
After each measurement, the flow cell was flushed with WFI. For particle identification, the following
settings were used: 3 µm distance to the nearest neighbor, particle segmentation thresholds for dark
pixels and light pixels of 13 and 10, respectively. The particle size was reported as the equivalent
spherical diameter (ESD). Frames were collected with VisualSpreadsheet® 4.7.6 software and were
evaluated for total particle counts of cumulative particles greater or equal to 1 µm, 10 µm and, 25 µm
per mL.
3. Results
3.1. Solid State
3.1.1. Residual Moisture Content and Specific Surface Area

The results for residual moisture (RM) and specific surface area (SSA) determination are presented
in Figure 2. For the low concentrated mAb formulation with sucrose (F1), three different storage
temperatures are shown (Figure 2a). Directly after freeze-drying, the SSA results revealed identical
values for CFD, 0.63 m2 /g ± 0.02 m2 /g, and MFD, 0.66 m2 /g ± 0.05 m2 /g. Irrespective of the storage
temperature and time point, neither differences nor relevant changes in specific surface areas were
observed. With regard to residual moisture content, analysis showed low values for both CFD and
MFD (1.1% ± 0.1% and 1.0% ± 0.5%, respectively). These values changed only slightly over 24 weeks
at all investigated storage temperatures. However, within the microwave-processed products, some
samples exhibited higher variances represented by higher standard deviations, which were not found
Pharmaceutics
within 2020, 12, x FOR PEER
the conventionally REVIEW samples.
freeze-dried 8 of 21
Figure 2. Specific
Figure 2. Specific surface
surfacearea
area(bars)
(bars)and
andresidual
residualmoisture
moisturecontent
content(squares)
(squares)results over
results thethe
over course of
course
24 weeks
of 24 of of
weeks storage at at
storage the respective
the respective storage
storagetemperature
temperaturefor
for(a)
(a)F1,
F1,(b)
(b)F2,
F2,(c)
(c)F3,
F3,and
and(d)
(d)F4.
F4. Values
Values
shown
shown represent
represent the
the mean
mean value (n == 3)
value (n 3) ± standard deviation.
± standard deviation.
Similar results were found for the low concentrated mAb formulation stabilized with trehalose
Similar results were found for the low concentrated mAb formulation stabilized with trehalose
(Figure 2b). No relevant differences or changes were observed for the specific surface area over the
(Figure 2b). No relevant differences or changes were observed for the specific surface area over the
course of 24 weeks storage at refrigerator temperature or 40 ◦ C. Mean values almost remained at initial
course of 24 weeks storage at refrigerator temperature or 40 °C. Mean values almost remained at
values of 1.27 m2 /g ± 0.01 m2 /g and 1.22 m2 /g ± 0.06 m2 /g for CFD and MFD, respectively. In regards
initial values of 1.27 m²/g ± 0.01 m²/g and 1.22 m²/g ± 0.06 m²/g for CFD and MFD, respectively. In
to residual moisture, MFD samples appeared to be moister than CFD samples (1.2% ± 0.8% vs. 0.3% ±
regards to residual moisture, MFD samples appeared to be moister than CFD samples (1.2% ± 0.8%
0.0%). These differences remained over the course of six months of storage irrespective of the storage
vs. 0.3% ± 0.0%). These differences remained over the course of six months of storage irrespective of
◦ C (0.7% ± 0.0%), unlike MFD cakes.
temperature. Yet, the moisture
the storage temperature. content
Yet, the in CFD
moisture cakesindoubled
content CFD cakesat 40doubled at 40°C (0.7% ± 0.0%), unlike
However, high variances within MFD samples may have masked
MFD cakes. However, high variances within MFD samples may have masked such effects. such effects.
Unlike
Unlike thethesucrose
sucrose(F1) (F1)and
andtrehalose
trehalose(F2) formulations,
(F2) formulations, lowlow
concentration
concentrationmAbmAb formulations with
formulations
arginine phosphate (F3) exhibited differences with respect
with arginine phosphate (F3) exhibited differences with respect to specificto specific surface area (Figure 2c). Initial
surface area (Figure 2c).
measurements directly after freeze-drying revealed values of 1.33 m 2 /g ± 0.09 m2 /g and 0.95 m2 /g
Initial measurements directly after freeze-drying revealed values of 1.33 m²/g ± 0.09 m²/g and 0.95
± 0.04±m 2 /g for CFD and MFD, respectively. Slight, but non-significant changes over storage were
m²/g 0.04 m²/g for CFD and MFD, respectively. Slight, but non-significant changes over storage
observed.
were observed.The initially different
The initially SSA values
different SSA correlated inverselyinversely
values correlated with the withobserved residual moisture
the observed residual
mean values (CFD: 1.0% ± 0.1% and MFD: 2.8% ± 0.4%). A micro-collapse
moisture mean values (CFD: 1.0% ± 0.1% and MFD: 2.8% ± 0.4%). A micro-collapse within MFD within MFD samples
was assumed.
samples was assumed.
F4,
F4, which was
which was comprised
comprised of of aa 1:1-mixture
1:1-mixture (weight-wise)
(weight-wise) of of sucrose
sucrose andand mAb2,
mAb2, was was dried
dried with
with aa
different
different microwave-setup.
microwave-setup. By By this,
this, high
high variances
variances inin residual
residual moisture,
moisture, which
which occasionally
occasionally occurred
occurred
before
before within MFD samples, were not observed anymore (Figure 2d). For conventional freeze-dried
within MFD samples, were not observed anymore (Figure 2d). For conventional freeze-dried
samples, ◦ C.
samples, mean
mean values
values changed
changed fromfrom 1.0%
1.0% (±0.0%)
(±0.0%) toto 1.2%
1.2% (±0.0%) over 24
(±0.0%) over 24 weeks
weeks ofof storage
storage at
at 40
40 °C.
Within
Within MFDMFD samples,
samples, aa similar
similar increase
increasefromfrom0.4%
0.4%(±0.0%),
(±0.0%), initially
initially to
to 0.6%
0.6% (±0.0%),
(±0.0%), was
was observed
observed
after six months at 40 ◦ C. In contrast, specific surface areas were found to remain unaffected by
after six months at 40 °C. In contrast, specific surface areas were found to remain unaffected by
accelerated storage conditions at values of 0.85 m²/g ± 0.15 m²/g and 0.89 m²/g ± 0.04 m²/g for CFD
and MFD, respectively.
3.1.2. X-Ray Powder Diffraction (XRD)

accelerated storage conditions at values of 0.85 m2 /g ± 0.15 m2 /g and 0.89 m2 /g ± 0.04 m2 /g for CFD
and MFD, respectively.
3.1.2. X-Ray Powder Diffraction (XRD)

In order to confirm the amorphicity of all formulations, XRD was used. The results for the two
sucrose-based formulations, F1 and F4, are presented in Figure S1 and were directly compared to the
pure excipient sucrose. No indications of crystallization were found.
Amorphous halos and the absence of typical peaks [29,30] were found for trehalose-based
formulations (Figure S2).
An overall XRD-amorphicity, represented by an amorphous halo, was also found for the
significantly moister arginine phosphate formulations (Figure S3). A reference diffractogram for
recrystallized arginine phosphate was derived by intentionally exposing one MFD vial to a moist
atmosphere overnight.
3.2. Protein-Related Quality Attributes
3.2.1. Reconstitution and Subvisible Particles (SvP)

Before liquid analysis, lyophilized products needed to be reconstituted. Within one formulation,
no significant difference between the distinct drying protocols was seen. However, small differences
with regard to other formulation were observed (Table 3).
Table 3. Reconstitution times of the different formulations.
Formulation Reconstitution Time (s)

F1 ≤30
F2 ≤30
F3 ≤50
F4 ≤120
The subvisible particle counts (SvP) obtained by light obscuration for F1 are presented in
Figure 3a,b. All vials analyzed originated from the same filtered bulk formulation, which is why initial
particle counts were the same before a certain drying or storage scheme was applied. Light obscuration
measurements revealed relatively low particle counts per mL of 2212 ± 565, 26 ± 2 ,and 4 ± 3 for
≥1 µm, ≥10 µm, and, ≥25 µm, respectively. An increase of +94% (4300 ± 546) and +49% (3303 ± 651)
for cumulative particles ≥1 µm/mL was observed directly after freeze-drying. In Figure 3a the results
after storage over 24 weeks at 4 ◦ C and 25 ◦ C are shown. After six months at refrigerator temperature,
particle counts were stabilized close to values prior to freeze-drying of 2678 ± 307 (CFD) and 2227 ±
225 for particles ≥1 µm/mL. At 25◦ C, storage temperature, subvisible particle counts were only slightly
elevated for CFD (2953 ± 295), but not for MFD (1937 ± 247). No increase in bigger particles, ≥10 µm,
and ≥25 µm, was seen at any of the storage conditions. Figure 3b shows that storage over 24 weeks at
40 ◦ C caused an increase by factor 2.3 (5106 ± 237) for conventionally FD, and an increase of 36% (3003
± 1058) for microwave-assisted FD, in ≥1 µm particles. However, bigger variances were found for
MFD samples.
and 2227 ± 225 for particles ≥1 µm/mL. At 25°C, storage temperature, subvisible particle counts were
only slightly elevated for CFD (2953 ± 295), but not for MFD (1937 ± 247). No increase in bigger
particles, ≥10 µm, and ≥25 µm, was seen at any of the storage conditions. Figure 3b shows that storage
over 24 weeks at 40 °C caused an increase by factor 2.3 (5106 ± 237) for conventionally FD, and an
increase of 36% (3003 ± 1058) for microwave-assisted FD, in ≥1 µm particles. However, bigger
variances were found for MFD samples.
Pharmaceutics
Figure 2020, 12, x FOR
3. Subvisible PEER REVIEW
particle (SvP) counts for formulation F1 measured by light obscuration 10 of
and21
size-exclusion chromatography results. The bar charts represent the subvisible particle counts for the
Figure 3. Subvisible particle (SvP) counts for formulation F1 measured by light obscuration and size-
respective storage temperatures (a) 4 ◦ C, 25 ◦ C, and (b) 40 ◦ C. Bars represent the mean value ± standard
exclusion chromatography results. The bar charts represent the subvisible particle counts for the
deviation for three
respective individual
storage vials. (a) 4 °C, 25 °C, and (b) 40 °C. Bars represent the mean value ±
temperatures
standard deviation for three individual vials.
Figure 4 represents the SvP counts analyzed by flow-imaging microscopy for F2. The same filtered
bulk formulation
Figure 4was used for
represents theallSvP
vials analyzed.
counts analyzedInitially, relatively low
by flow-imaging particle counts
microscopy for F2. per
ThemL of 1867
same
± 1784, 90 ± bulk
filtered 20 and 20 ± 35 for
formulation ≥1used
was ≥10allµm,
µm, for and
vials ≥25 µm,
analyzed. respectively,
Initially, were
relatively lowfound. Acounts
particle slight per
increase
by 24%mL(2320
of 1867
± ±599)
1784,
and90 37%
± 20 and
(2549 20 ±
± 35 forfor
677) ≥1 µm, ≥10 µm, particles
cumulative and ≥25 µm, respectively,
≥1µm/mL waswere found.directly
observed A
slight increase by 24% ◦ (2320 ± 599) and 37% (2549 ± 677) for cumulative
after freeze-drying. At 4 C storage temperature, particle numbers ≥1 µm/mL settled around initialparticles ≥1µm/mL was
valuesobserved
after 24directly
weeks, butafterthe
freeze-drying.
cumulativeAt 4 °Cofstorage
count bigger temperature, particle(Figure
particles increased numbers ≥1However,
4a). µm/mL no
settled around initial values after 24 weeks, but the cumulative count of bigger particles increased
significant changes were observed. A dramatic increase for SvP ≥1 µm/mL was found over the course
(Figure 4a). However, no significant changes were observed. A dramatic increase for SvP ≥1 µm/mL
of six months at 40◦ C for both CFD (37909 ± 4337) and MFD (18947 ± 6753), as shown in Figure 4b. The
was found over the course of six months at 40°C for both CFD (37909 ± 4337) and MFD (18947 ± 6753),
meanasSvP count
shown values4b.
in Figure ≥10mean
forThe µm SvP ≥25 µm
and count alsofor
values increased
≥10 µm anddrastically, even
≥25 µm also thoughdrastically,
increased high standard
deviations lowered
even though highthe significance.
standard deviations However,
lowered the ansignificance.
upward trend couldanbe
However, assumed
upward trendfor subvisible
could be
particles ≥10 µm/mL,
assumed in conventionally
for subvisible processed
particles ≥10 µm/mL, samples. processed samples.
in conventionally
Figure 4. Subvisible
Figure particle
4. Subvisible (SvP)(SvP)
particle counts for trehalose-based
counts formulation
for trehalose-based F2 measured
formulation by flow-imaging
F2 measured by flow-
imaging microscopy.
microscopy. Barrepresents
Bar chart (a) chart (a) represents
the SvP the SvP counts
counts at refrigerator
at refrigerator storage
storage (b)(b)atat4040◦°C.
andand C. Bars
Bars represent
represent the meanthe mean
value ± value ± standard
standard deviation
deviation for three
for three individual
individual vials.
vials.
The SvP
The SvP counts
counts for the
for the lowlow concentration
concentration mAbmAbformulation
formulationstabilized
stabilized by
by arginine
argininephosphate,
phosphate, F3,
F3, are shown in Figure 5. Prior to freeze-drying, relatively low particle
are shown in Figure 5. Prior to freeze-drying, relatively low particle counts percounts per mLmL
of 1991 ± 1490,
of 1991 ± 1490,
68 ± 42 and 20 ± 21 for ≥1 µm, ≥10 µm, and ≥25 µm were found, respectively. At 4 °C (Figure ◦ 5a)
68 ± 42 and 20 ± 21 for ≥1 µm, ≥10 µm, and ≥25 µm were found, respectively. At 4 C (Figure 5a)
storage temperature, a small increase in mean values for cumulative particles ≥1 µm was found over
storage temperature, a small increase in mean values for cumulative particles ≥1 µm was found over
time, although particle counts for this size category settled around the initial values. For particles ≥10
time,µm
although
and ≥25particle countsstronger
µm, a slightly for thisincrease
size category
in meansettled around
values was the initial
observed, values.
although For particles
vast standard
deviations lowered the significance.
≥10 µm and ≥25 µm, a slightly stronger increase in mean values was observed, although vast standard
deviations lowered
Pharmaceutics the significance.
2020, 12, x FOR PEER REVIEW 11 of 21
Figure 5. Subvisible particle (SvP) counts for the arginine phosphate formulation F3 measured by
Subvisible particle
Figure 5. Subvisible particle (SvP)
(SvP) counts
counts for
for the
the arginine
arginine phosphate
phosphate formulation
formulation F3
F3 measured
measured by
by
flow-imaging microscopy. Bar chart (a) represents the SvP counts at refrigerator storage and (b) at 40
flow-imaging microscopy. Bar chart (a) represents the SvP counts at refrigerator storage and (b) at 40
°C.
◦ C. Bars represent the mean value ± standard deviation for three individual vials.
°C. Bars represent the mean value ±± standard
Bars represent standard deviation
deviation for
for three
three individual
individual vials.
vials.
At accelerated
At accelerated storage
acceleratedstorage conditions
storageconditions
conditions (Figure
(Figure 5b), a moderate
5b), a moderate increase in all size
increase categories was seen.
At (Figure 5b), a moderate increase in allinsize
all categories
size categories was
was seen.
Especially
seen. after
Especially six
after months
six monthsstorage, the
storage, conventionally
the conventionally freeze-dried
freeze-dried sample
sample showed
showed aa significant
significant
Especially after six months storage, the conventionally freeze-dried sample showed a significant
increase for ≥1 µm, ≥10 µm, and ≥25 µm with particle counts per
per mL of of 5913±± 1584
1584 (3× higher), 394 ±
increase for ≥1
increase for µm, ≥10
≥1 µm, ≥10 µm, and
µm, ≥25 µm
and ≥25 with particle
µm with particle counts
counts per mL mL of 59135913 ± 1584 (3× 394 ±
higher), 394
(3× higher), ±
97
97 (6×
(6× higher),
higher), and
and 78
78 ±
± 21
21 (4×
(4× higher),
higher), respectively.
respectively. In
In contrast,
contrast, the
the microwave-assisted
microwave-assisted freeze-dried
freeze-dried
97 (6× higher), and 78 ± 21 (4× higher), respectively. In contrast, the microwave-assisted freeze-dried
sample
sample at at the same
same conditionsshowed showed noincrease
increase for ≥1µm,
≥1≥1 µm, and only a modest increase by factor
sample at the
the sameconditions
conditions showedno no increasefor for µm,and and only a modest
only a modest increase
increaseby by
factor 2.6
factor
2.6
and and
2.2 2.2
for for cumulative
cumulative particle
particle counts
counts ≥10 ≥10
µm µmand and
≥25 ≥25
µm, µm, respectively.
respectively.
2.6 and 2.2 for cumulative particle counts ≥10 µm and ≥25 µm, respectively.
For the formulation
For the with 50 g/L of mAb2 and only 5% (w/v), sucrose-stabilizer neither at
For theformulation
formulation with
with50 g/L
50 of
g/LmAb2 and only
of mAb2 and5%only (w/v),5%sucrose-stabilizer neither at refrigerator
(w/v), sucrose-stabilizer neither at
refrigerator
(Figure 6a) nor (Figure ◦
at 40 6a)6a) nor
C storage at 40 °C storage temperature (Figure 6b), was a significant change in
refrigerator (Figure nor at temperature
40 °C storage (Figure 6b), was (Figure
temperature a significant 6b), change in subvisible
was a significant particles
change in
subvisible particles
observed. particles
The initial observed.
formulationThe initial
before formulation
FD revealed before
low FDFD revealed
particle countslow particle counts per mL
± 1153,
subvisible observed. The initial formulation before revealed lowper mL of
particle 2051 per
counts mL
of 2051 ±
± 41±and 1153,7± 40 ± 41 and 7 ± 6 for ≥1 µm, ≥10 µm, and ≥25 µm, respectively. The results obtained
40 2051
of 1153, 406±for 41 ≥1andµm,7 ± 6≥10forµm,
≥1 µm,and≥10≥25µm, µm,and respectively. The resultsThe
≥25 µm, respectively. obtained
resultsshowed
obtaineda
showed
larger a largerbyincrease
increase 74% by ±
(3579 74% (3579
2243) for ±CFD
2243)andfor25%CFD and ±25%
(2555 97) (2555
for ± 97)infor
MFD, MFD,to
regards incumulative
regards to
showed a larger increase by 74% (3579 ± 2243) for CFD and 25% (2555 ± 97) for MFD, in regards to
cumulative particles
particles ≥1 particlesat four ≥1 µm at four degrees celsius storage over six months, compared to
◦ C.storage at
cumulative µm ≥1degrees
µm at fourcelsius storagecelsius
degrees over six months,
storage over compared
six months, to storage
compared at 40 to Yet, at all
storage at
40 °C. Yet,
conditions at all conditions
observed, observed,
larger particle larger
categories, particle
i.e., ≥10 categories,
µm and ≥25 i.e., ≥10 µm and ≥25 µm, revealed an
40 °C. Yet, at all conditions observed, larger particle categories, ≥10 revealed
i.e.,µm, µm and an ≥25increased
µm, revealednumber an
increased
of particle number
counts. of particle due
However, counts.
to However,
high standard due to high standard
deviations within adeviations
sample, no within a sample,
significant changesno
increased number of particle counts. However, due to high standard deviations within a sample, no
significant changes
were detectable. were detectable.
significant changes were detectable.
Figure 6. Subvisible
Subvisible particle
particle (SvP)
(SvP) counts
counts for
for the
the high
high concentration
concentration mAb
mAb formulation
formulation with
with 50
50 g/L
Figure 6. Subvisible particle (SvP) counts for the high concentration mAb formulation with 50 g/L
stabilized by sucrose, measured by flow-imaging microscopy. Bar chart (a) represents the SvP counts
stabilized by sucrose, measured by flow-imaging microscopy. Bar chart (a) represents the SvP counts
at refrigerator storage and (b) at 40 ◦°C.
C. Bars represent the mean value ±± standard deviation for three
at refrigerator storage and (b) at 40 °C. Bars represent the mean value ± standard deviation for three
individual vials.
individual vials.
3.2.2. Weak Cation Exchange Chromatography (CEX)

In order to generically quantify different protein degradation pathways, e.g., deamidation, a
salt-gradient weak cation exchange chromatography was used. The CEX data for F1 was not
salt-gradient weak cation exchange chromatography was used. The CEX data for F1 was not

salt-gradient weak
collected. But forcation exchange chromatography
the trehalose-based was used.
low concentration mAb The CEX data for
formulation F1 data
(F2), was not
fromcollected.
CEX
Butmeasurements
for the trehalose-based
is shown inlow concentration
Figure 7a,b for themAb formulation
respective storage(F2), data fromDirectly
temperatures. CEX measurements
after freeze- is
shown in Figure
drying, 7a,b for
irrespective the applied
of the respective storage
drying temperatures.
protocol Directly
(CFD or MFD), after freeze-drying,
relative amounts of theirrespective
different
of species
the applied drying protocol (CFD or MFD), relative amounts of the different
were found to be alike. For storage at refrigerator temperature, (Figure species werespecies
7a) acidic found to
be dropped
alike. Forbystorage
roughlyatone
refrigerator
percent fortemperature,
both drying (Figure 7a)whereas
protocols, acidic species dropped
basic species by roughly
slightly increasedone
percent for Somewhat
by 1.5%. both drying protocols,
more whereas
pronounced basicwere
changes species slightly
observed forincreased
storage at by 1.5%.
40 °C Somewhat
(Figure more
7b). While
acidic species
pronounced increased
changes were by three percent,
observed for storage 40 ◦ C (Figure
basicatspecies rose by7b).
4.7% and 3.5%
While acidicfor CFD and
species MFD, by
increased
respectively.
three percent, basic species rose by 4.7% and 3.5% for CFD and MFD, respectively.
Figure
Figure7. 7.Relative amountofofacidic
Relative amount acidic
andand basic
basic species
species obtained
obtained by high-performance
by high-performance (HP)-weak
(HP)-weak cation
cation exchange chromatography for trehalose-based formulation ◦ C and (b) 40 ◦ C storage
exchange chromatography for trehalose-based formulation F2 atF2(a)at 4(a)°C4 and (b) 40 °C storage
temperature.
temperature. InIn
(c),(c),
the relative
the relativepercentages
percentagesofofmonomer
monomerand andhigh
highmolecular
molecular weightweight species
species (HMW) at
theatrespective storage
the respective temperature
storage temperatureover storage
over storagetime
timegained
gained by HP-size
HP-sizeexclusion
exclusionchromatography
chromatography
(SEC)
(SEC)analysis
analysisare
are presented.
presented.
ForFor
thethe
arginine phosphate-formulation
arginine phosphate-formulation (F3),(F3),
CEXCEXdata data
is shown in Figure
is shown 8a,b. A8a,b.
in Figure smallAdifference
small
of difference
1.7% in theofinitial relative
1.7% in amount
the initial of acidic
relative amount species wasspecies
of acidic found,was which equalized
found, for the two
which equalized fordrying
the
two drying
protocols overprotocols
24 weeksover 24 weeks
storage 4 ◦ C (Figure
time atstorage time at8a).
4 °CAt(Figure
the same8a).conditions,
At the same conditions,
basic basic
species slightly
species for
increased slightly
bothincreased for both
CFD (+1.8%) and CFD
MFD(+1.8%)
(+1.3%),and MFD (+1.3%),
revealing a smallrevealing
deviation. a small deviation.
Acidic Acidic a
species showed
species showed a noticeable change by 15.9% (CFD) and ◦26.6% (MFD) at 40 °C storage
noticeable change by 15.9% (CFD) and 26.6% (MFD) at 40 C storage temperature (Figure 8b), whereas temperature
(Figure
basic 8b),remained
species whereas basic species
almost remained
the same with almost
changes theby
same
2.5%with
(CFD)changes by 2.5%
and −0.5% (CFD) and −0.5%
(MFD).
(MFD).
Figure8. 8. Relative
Figure Relative amount
amount of of acidic and basic
acidic and basic species
species obtained
obtainedby byHP-weak
HP-weakcationcationexchange
exchange
chromatography ◦ ◦
chromatographyfor forarginine
arginine phosphate-based formulationF3
phosphate-based formulation F3atat(a)
(a)4 4°CCand
and(b)(b)4040°C C storage
storage
temperature.
temperature. InIn
(c), the
(c), therelative
relativepercentages
monomerand andhigh
highmolecular
molecular weight
weight species (HMW) at
species (HMW)
theatrespective storage
the respective temperature
storage temperature over storage
over time
storage gained
time gainedbyby
HP-SEC
HP-SECanalysis
analysisare
arepresented.
presented.
InIn
the
thecase
caseofofthe
thehigh
highconcentration
concentration mAbmAb formulation stabilizedwith
formulation stabilized withsucrose
sucrose (F4),
(F4), almost
almost nono
differences were found between conventional
differences were found between conventional and microwave-assisted freeze-dried products
microwave-assisted freeze-dried products with with
regard to to
regard CEX
CEX results
results(Figure
(Figure9a,b).
9a,b).At
At refrigerator temperature(Figure
refrigerator temperature (Figure9a),
9a),changes
changes
inin both
both species
species
and both
and bothdrying
dryingprotocols
protocolsranged
rangedwithin
within less
less than 0.5%.
0.5%. While
Whilesimilar
similarobservations
observations were
were made
made forfor
thethe acidic
acidic speciesatat4040◦ C,
species °C,basic
basic species
species slightly
slightly rose
rose by
by2.7%
2.7%and
and3%3%for
forCFD
CFDand
andMFD
MFD samples,
samples,
respectively.
respectively. However,nonodifference
However, differencebetween
betweenthe thedrying
drying procedures
procedures was
wasobserved.
observed.
Figure
Figure 9. Relative amount
9. Relative amount of of acidic
acidic and
and basic
basic species
species obtained
obtained by
by HP-weak
HP-weak cation
cation exchange
exchange
chromatography for arginine phosphate-based formulation F3 at (a) 4 ◦ C and (b) 40 ◦ C storage
chromatography for arginine phosphate-based formulation F3 at (a) 4 °C and (b) 40 °C storage
temperature.
temperature. In
In (c),
(c), the
the relative
relativepercentages
monomerand andhigh
highmolecular
molecular weight species
weight (HMW)
species (HMW)at
the respective storage temperature over storage time gained by HP-SEC analysis are presented.
at the respective storage temperature over storage time gained by HP-SEC analysis are presented.
3.2.3. Size Exclusion Chromatography (SEC)
3.2.3. Size Exclusion Chromatography (SEC)
The relative amount of monomeric and high molecular weight (HMW) species was assessed by
The relative amount of monomeric and high molecular weight (HMW) species was assessed by
high-performance size-exclusion chromatography. The results of the HP-SEC analysis for the low
high-performance size-exclusion chromatography. The results of the HP-SEC analysis for the low
concentration mAb formulation with 10% (w/v) sucrose (F1) are displayed in Figure S4. Irrespective
concentration mAb formulation with 10% (w/v) sucrose (F1) are displayed in Figure S4. Irrespective
of the storage temperature, no changes in monomer content occurred. In other words, the relative
of the storage temperature, no changes in monomer content occurred. In other words, the relative
amount of monomeric species ranged between 99.0% to 99.1% at all analyzed time points, and all
amount of monomeric species ranged between 99.0% to 99.1% at all analyzed time points, and all
investigated storage temperatures.
investigated storage temperatures.
Only a slightly different picture was seen for the trehalose formulation F2 in Figure 7c. After
Only a slightly different picture was seen for the trehalose formulation F2 in Figure 7c. After six
six months at 4 ◦ C, the loss of monomer and the complementary rise in HMW was negligibly low,
months at 4 °C, the loss of monomer and the complementary rise in HMW was negligibly low,
basically within the range of SEC sample standard deviation. A small decrease by −0.6% (98.2% ±
basically within the range of SEC sample standard deviation. A small decrease by −0.6% (98.2% ±
0.1%) for CFD and by −0.4% (98.4% ± 0.1%) for microwave-assisted freeze-dried lyophilizates was
0.1%) for CFD and by −0.4% (98.4% ± 0.1%) for microwave-assisted freeze-dried lyophilizates was
observed in accelerated storage conditions of 40 ◦ C for 24 weeks.
observed in accelerated storage conditions of 40 °C for 24 weeks.
In arginine phosphate-based monoclonal antibody formulation (F3), more changes were observed
In arginine phosphate-based monoclonal antibody formulation (F3), more changes were
(Figure 8c). At refrigerator temperature, a small decrease by less than one percent in monomer content,
observed (Figure 8c). At refrigerator temperature, a small decrease by less than one percent in
and thus, an increase in HMW aggregates to less than 2% overall HMW species, occurred. However,
monomer content, and thus, an increase in HMW aggregates to less than 2% overall HMW species,
a significant loss in monomeric content down to 95.7% ± 0.4% and 96.8% ± 0.3% for CFD and MFD,
occurred. However, a significant loss in monomeric content down to 95.7% ± 0.4% and 96.8% ± 0.3%
respectively, was seen. This was counterbalanced by an increase in high molecular weight species to
for CFD and MFD, respectively, was seen. This was counterbalanced by an increase in high molecular
4.3% ± 0.4% and 3.2% ± 0.3% for conventionally and microwave-assisted freeze-dried samples.
weight species to 4.3% ± 0.4% and 3.2% ± 0.3% for conventionally and microwave-assisted freeze-
The formulation with 50 g/L of mAb2 (F4) showed a higher monomeric SEC-purity of 99.9%
dried samples.
± 0.0%, compared to mAb1 formulations prior to freeze-drying (Figure 9c). At refrigerator storage
The formulation with 50 g/L of mAb2 (F4) showed a higher monomeric SEC-purity of 99.9% ±
temperatures, only a small loss of 0.2% relative monomer content was observable regardless of the
0.0%, compared to mAb1 formulations prior to freeze-drying (Figure 9c). At refrigerator storage
temperatures, only a small loss of 0.2% relative monomer content was observable regardless of the
initially used drying procedure. Even for 24 weeks at 40 °C, the monomer content for CFD and MFD
initially used drying procedure. Even for 24 weeks at 40 ◦ C, the monomer content for CFD and
MFD samples decreased only slightly to 98.5% ± 0.0% and 98.2% ± 0.0%, respectively. This loss
was compensated by an increase in HMW to 1.5% ± 0.0% and 1.8% ± 0.0% for conventionally and
microwave-assisted freeze-dried samples, respectively.
4. Discussion
4.1. Stability with Regard to Solid State Properties

In the case of freeze-dried products, attributes like residual moisture content, solid-state, and
specific surface area are critical to assess and to monitor over storage and shelf-life [31]. With respect to
the specific surface area, the low mAb concentration formulations with sucrose and trehalose, F1 and F2
(Table 1), respectively, showed identical values for each formulation for the two distinct drying protocols
directly after freeze-drying, and showed no change over the duration of storage. This strongly indicated
the absence of a microscopic collapse, which means that the initially determined ice and successive pore
structure remained during the two different drying protocols for respective formulations [32]. Regarding
the residual moisture content similar mean values were found for the sucrose-based formulation F1, at
1.1% ± 0.1% and 1.0% ± 0.5% for CFD and MFD, respectively (Figure 2a). This moisture level was
kept over the storage duration, even at elevated temperatures. However, high variances in some MFD
samples were observed. Such high variances among samples of one microwave-batch may have been
caused by non-uniform temperature distribution. This non-uniformity in microwave heating was
reported in the literature to be one major challenge associated with that drying technique [15,18,33].
The resulting appearance of cold and hot spots was described as multifactorial and may be dependent
on the chosen mode (multimode, single-mode) [15,34], oven design [33], sample composition, geometry
of the frozen good [15,35–37], occurrence of standing wave effect [38], and drying duration [39].
In contrast, in trehalose samples (F2), residual moisture levels differ already after the freeze-drying
step, i.e., 0.3% ± 0.0% for conventionally dried and 1.2% ± 0.8% for microwave-assisted dried samples
(Figure 2b). The CFD samples stored at 40 ◦ C showed an increase to 0.7% ± 0.0%. This was most likely
due to moisture uptake of the cake from the rubber stopper, as was described by Pikal and Shah [40].
The equilibration of stopper moisture and cake moisture was found to be kinetically dependent on
storage temperature in the first place. Because of the high variances in MFD samples, such an effect
may have happened but could not be detected.
An arginine phosphate-based formulation (F3) was expected to be different. Firstly, because
of the permanently charged character of arginine which causes high dielectric loss, i.e., the ability
of a material to absorb electromagnetic energy [15], as reported by Meng et al. [41]. Of course, the
dielectric properties of a material may vary depending on the exact composition, density, temperature,
and frequency [15]. Nevertheless, different behavior of such formulation within the electromagnetic
microwave field was expected.
Secondly, because of the reported complex molecular structure of arginine phosphate [42].
Although the reported structure was described for the crystalline state, similar intense interactions in
the glassy state were assumed [43,44]. For formulation F3 specific surface area values were reduced
by 39% to 0.95 m2 /g ± 0.04 m2 /g for MFD, compared to 1.33 m2 /g ± 0.09 m2 /g for CFD, initially after
application of the different drying protocols (Figure 2c). It is assumed that this shift in the SSA is
associated with a microscopic collapse within the amorphous matrix, which may have been favored by
the permanently charged matrix leading to enhanced absorption of microwave energy [41]. As the
desorption step of the unfrozen water is highly SSA-dependent [45], the micro-collapse seems likely to
be the cause for higher mean residual moisture in MFD (2.8% ± 0.4%), compared to CFD (1.0% ± 0.1%).
Over the course of storage, no significant change in glassy state properties was observed.
The 50 g/L mAb2 formulation stabilized with sucrose (F4) appeared to be different with respect to
residual moisture content (Figure 2d). The occasionally emerging high variances within one sample
(n = 3) produced by MFD was not observed anymore. Two possible reasons are: 1) The formulation,
which consisted of roughly 50% monoclonal antibody, may have changed the matrix properties; and 2)
the newly implemented microwave technical setup as described in the materials and method section.
The authors assume that the increase in batch homogeneity within MFD samples was mainly due to
the change in machinery setup. This is believed because: Firstly, similar formulations with the same
mAb have been dried with the previously described MFD setup [20], and they revealed the same high
deviations that have been observed with the low concentrated mAb formulations F1–F3 in this study
(Table S1).
Secondly, with typical sample sizes of n = 3, potential batch inhomogeneities were observable.
Thirdly, the solid-state semiconductor setup led to a more stable and better tuneable power input
during the course of drying. Supportively, Bianchi et al. [46] simulated the physical behavior of apple
slices under microwave-assisted vacuum drying processing comparing magnetron and solid-state
technology. They concluded that with the latter, an improved heating pattern uniformity could
be achieved.
With regard to X-ray diffraction analysis, all formulations revealed an XRD-amorphous solid-state
exhibiting an amorphous halo (Figures S1–S3). By that, the authors suppose no adverse effect of
microwaves on the crystallization tendency of the investigated matrices. Even moister samples have
not revealed any indication of recrystallization.
4.2. Stability with Regard to Protein-Related Properties

The sucrose-based formulation with 5 g/L of mAb1 (F1) showed no clear trend in subvisible
particles (LO) at 4 ◦ C and 25 ◦ C storage temperature (Figure 3a) and no difference between CFD and
MFD either. Only directly after the freeze-drying procedure an increase of 94% (4300 ± 546) and
49% (3303 ± 651) for cumulative particles ≥1 µm/mL for CFD and MFD, respectively, was observed.
Figure 3b shows that storage over 24 weeks at 40 ◦ C caused an increase by factor 2.3 (5106 ± 237) for
conventionally FD and an increase of 36% (3003 ± 1058) for microwave-assisted FD in ≥1 µm particles.
It was expected that this slight change in SvP at accelerated storage conditions might correlate with an
increase in the relative amount of high molecular weight species assessed by HP-SEC. However, no
such effect on SEC data was seen (Figure S4). Across all storage temperatures and regardless of the
drying protocol used, no change in neither soluble aggregates nor loss of monomer was observed.
For the trehalose-based formulation (F2), the more sensitive flow-imaging microscopy technique
was used for subvisible particle determination (Figure 4). Initially, low particle counts per mL of 1867
± 1784, 90 ± 20 and 20 ± 35 for ≥1 µm, ≥10 µm, and ≥25 µm, respectively, were found, but yet bearing
unusually high variances. Directly after freeze-drying, most likely due to freeze-drying associated
stresses [47], a slight increase by 24% (2320 ± 599) and 37% (2549 ± 677) for cumulative particles
≥1 µm/mL was observed. No significant changes were observed for SvP at 4 ◦ C storage temperature
(Figure 4a). Taking chromatographic results into account, no HP-CEX (Figure 7a,b) or HP-SEC
(Figure 7c) changes were observed, emphasizing sufficient and comparable stabilization in both drying
populations at 4 ◦ C. In contrary, a dramatic increase for subvisible particles ≥1 µm/mL was found over
the course of six months at 40 ◦ C, for both CFD (37,909 ± 4337) and MFD (18,947 ± 6753), as shown in
Figure 4b. A similarly pronounced increase was also for bigger particles observable, although high
standard deviations lowered significance. Yet, an upward trend was assumed for subvisible particles
≥10 µm/mL in conventionally processed samples. With respect to chemical degradation (Figure 7b), a
linear increase with shallow slope was found for both acidic and basic species irrespective of the drying
procedure, although basic species increased slightly more in CFD samples. An increase in the different
species could be related to several different changes within the protein molecule depending on primary
structure, cell line, formulation conditions, and so forth [48–50]. Reviewed by Du et al. [48] in 2012,
a major contribution to acidic species was associated with the deamidation reaction of asparagine
residues. For basic species, depending on primary structure, they discussed different causes covering
C-terminal basic amino residues, including incomplete cyclization of the N-terminal, but also the
formation of aggregates. It could be suspected that the stronger increase in high molecular weight
aggregates (Figure 7c) at 40 ◦ C in CFD samples is related to the stronger rise of basic species in HP-CEX
data. However, no follow-up investigation by (partial) protein digestion or by LC-MS was conducted.
Nonetheless, as differences between conventionally dried and microwave-assisted freeze-dried samples
were rather marginal, comparable stability in the trehalose formulation was deduced.
Arginine phosphate, as discussed in Section 4.1, was expected to be an exceptional formulation,
especially challenging when drying with electromagnetic waves. In a recently published review by
Stärtzel [51], several examples of successful protein stabilization by arginine salts in the glassy state
were shown. Within our study, we found only tiny changes at refrigerator temperature in regards to
subvisible particles (Figure 5a). Primarily the mean values for bigger sized particles (≥10 µm and
≥25) increased after 24 weeks of storage, yet were not overly due to higher variances. HP-CEX data
(Figure 8a) and HP-SEC data (Figure 8c) supported this. Only a slight degradation shown by relative
monomer loss of less than 1% in both CFD and MFD, and a small increase in basic species of less
than 2%, were found at 4 ◦ C. In accelerated storage conditions (Figure 5b), a moderate increase in
all size categories was seen. The conventionally freeze-dried sample especially showed a significant
increase for cumulative particle counts ≥1 µm, ≥10 µm and, ≥25 µm, with factors of 3× to 6× higher
counts after six months of storage. In contrast, the microwave-assisted freeze-dried sample at the same
conditions showed only a moderate increase by bigger particles ≥10 µm and ≥25 µm. Connecting
that to the HP-CEX data (Figure 8b), a contrary picture can be drawn. One the one hand, a noticeable
change in acidic species by 15.9% (CFD) and 26.6% (MFD) at 40 ◦ C storage temperature was found,
whereas basic species remained almost the same for both drying procedures. On the other hand,
HP-SEC data (Figure 8c) revealed a loss in monomeric content down to 95.7% ± 0.4% and 96.8% ± 0.3%
for CFD and MFD, respectively. Taking the above into account, it appeared that the conventionally
dried sample showed a higher degree of degradation, exhibiting a higher increase in high molecular
weight aggregates, accompanied by an increased particle count in flow-imaging microscopy. However,
HP-CEX results indicated that a moderate portion of the main charge variant underwent chemical
reactions leading to significantly increased acidic species, being more pronounced in MFD samples.
Stärtzel et al. [44] investigated different arginine salts on their stabilizing potential in the glassy state.
For that, they also calculated the relaxation time τβ , “which may be regarded as proportional to the
inverse of molecular mobility for global motions” [52]. Subsequently, they related physical aggregation
rate constants at 40 ◦ C with the estimated ln(τβ ). They found that an arginine phosphate formulation
(64 g/L L-Arg, 16 g/L sucrose and 50 g/L mAb) revealed longer relaxation times than other arginine
formulations. These relaxation times unexpectedly had an inversely proportional correlation to the
observed aggregation constants, which suggested that increased molecular mobility had a positive
effect on protein stability [44]. This may be one explanation for the stability differences observed
between CFD and MFD. Microwave-assisted freeze-dried samples on average, revealed higher residual
moisture content (Figure 2c), compared to CFD. Residual water is known as a plasticizer of amorphous
matrices, which also consequently leads to increased molecular mobility [53], and therefore potentially
to reduced aggregation in an arginine-based system as reported by [44]. However, increased molecular
mobility may be associated with increased chemical degradation [54,55], giving a potential explanation
to the more distinct increase in acidic species MFD samples. Another explanation could be based on an
advantageous effect of the partial collapse in MFD samples, as it was observed by Schersch et al. [56],
for partially collapsed mannitol-sucrose formulations. All in all, microwave-assisted lyophilizates
with arginine phosphate, on the one hand, revealed an indication for more pronounced chemical
degradation, but on the other hand, showed a less severe increase in subvisible particles and aggregates.
The authors, therefore, conclude a comparable stability profile for CFD and MFD with reservations.
For the high concentration mAb2 formulation (F4), no clear trend could be derived from subvisible
particle analysis. Unexpectedly, particle counts at refrigerator temperature appeared to be higher
or at similar levels compared to 40 ◦ C after six months (Figure 6). With regard to HP-CEX results,
no difference between conventionally and microwave-assisted freeze-dried samples was observed
(Figure 9a,b). In size-exclusion chromatography, samples stored at 4 ◦ C exhibited a negligibly small
loss of monomer for both drying protocols. At accelerated conditions, a rather low loss in monomeric
species of 1.4% and 1.7% for CFD and MFD, respectively, was seen after six months of storage. For this
reason, the authors conclude that both sample populations derived from the two respective drying
protocols were comparable with respect to protein stability.
In the future, an improved prototype dryer with a sophisticated technical setup that provides the
operator with standard pharmaceutical freeze-drying features, such as freezing and stoppering within
the same machine, is needed. A combination of current pharmaceutical freeze-drying equipment with
modern semiconductor solid-state microwave generators is imaginable, in the authors’ opinion. For
future experiments, a look into relaxation behavior and potential differences between conventionally
freeze-dried and microwave-assisted freeze-dried solids could be of interest. Thermal history is
expected to be different. Moreover, a deeper look into potential chemical changes that may occur
during MFD should be taken. For those analytical techniques focusing on structural changes like FT-IR
and circular dichroism, but also methods like (peptide mapping) LC-MS, should be considered.
5. Conclusions
Microwave-assisted freeze-drying is an emerging technique recently introduced to the field
of pharmaceutical freeze-drying of biologicals [20,24]. Despite potentially huge time savings for
vial-based drying achievable by MFD [20], we were able to elucidate comparable stability profiles
for different monoclonal antibody formulations over storage times of 24 weeks. Although residual
moisture contents were found to be different between CFD and MFD, no adverse effect on protein
stability or crystallization tendency in matrices with higher residual moisture was found. Even the
occurrence of a microscopic collapse in the microwave-processed arginine phosphate mAb formulation
(F3) did not lead to decreased stability, with respect to solid state- and protein-related properties.
Moreover, our data indicate that with modern semiconductor solid-state microwave generators batch
homogeneities of microwave batches could be approximated to those of conventional freeze-drying.
However, the authors see a definite need for new machines complying with the requirements of
pharmaceutical manufacturing. The new generator setup presented may open up space for engineering
creativity to merge pharmaceutical needs with innovative heating techniques.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4923/11/12/674/s1,

Figure S1: Representative X-ray diffractograms of sucrose-based formulation F1 and F4 compared to unprocessed
pure crystalline sucrose (EMPROVE®exp sucrose) from the shelf. Top graphs represent conventionally freeze-dried
samples directly after FD (left) and after 24 weeks of storage at 40 ◦ C (right). The same applies to the bottom
graphs but with microwave-assisted freeze-dried samples. Figure S2: Representative X-ray diffractograms of
sucrose-based formulation F1 and F4 compared to unprocessed pure crystalline sucrose (EMPROVE®exp sucrose)
from the shelf. Top graphs represent conventionally freeze-dried samples directly after FD (left) and after 24
weeks of storage at 40 ◦ C (right). The same applies to the bottom graphs but with microwave-assisted freeze-dried
samples., Figure S3: Representative X-ray diffractograms of low concentration mAb formulation with arginine
phosphate (F3). The diffractogram of a MFD recrystallized sample of identical composition (gray line) was used for
identification of reference peaks. Top graphs represent conventionally freeze-dried samples directly after FD (left)
and after 24 weeks of storage at 40 ◦ C (right). The same applies to the bottom graphs but with microwave-assisted
freeze-dried samples, Figure S4: Relative percentages of monomer and high molecular weight species (HMW)
at the respective storage temperature over storage time gained by HP-SEC analysis are presented. Table S1:
Residual moisture content from two similar mAb2 formulations produced with the previous microwave setup [20]
compared to F4.
Author Contributions: Conceptualization, I.P. and G.W.; Data curation, J.H.G.; Formal analysis, J.H.G.; Funding
acquisition, G.W.; Investigation, J.H.G.; Methodology, J.H.G.; Project administration, J.H.G., Raimund Geidobler,
I.P. and G.W.; Resources, R.G., I.P., and G.W.; Supervision, R.G., I.P., and G.W.; Validation, J.H.G.; Visualization,
J.H.G.; Writing—Original draft, J.H.G.; Writing—Review & editing, R.G., I.P., and G.W.
Funding: This research was funded by Boehringer Ingelheim Pharma GmbH & Co. KG.
Acknowledgments: The support from the Global Technology Management of Boehringer Ingelheim Pharma
GmbH & Co. KG is kindly acknowledged. In addition, the authors thank Peter Püschner, Michael Eggers, and
Mirko Diers from Püschner GmbH & Co KG for technical support with the microwave vacuum dryer experiments.
Conflicts of Interest: The authors declare no conflict of interest. R.G. and I.P. are full-time employees of Boehringer
Ingelheim Pharma GmbH & Co. KG. The funders had no role in the design of the study; in the collection, analyses,
or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
References
1. Carpenter, J.F.; Pikal, M.J.; Chang, B.S.; Randolph, T.W. Rational Design of Stable Lyophilized Protein
Formulations: Some Practical Advice. Pharm. Res. 1997, 14, 969. [CrossRef] [PubMed]
2. Franks, F. Freeze-drying of bioproducts: Putting principles into practice. Eur. J. Pharm. Biopharm. 1998, 45,
221–229. [CrossRef]
3. Gervasi, V.; Dall Agnol, R.; Cullen, S.; McCoy, T.; Vucen, S.; Crean, A. Parenteral protein formulations: An
overview of approved products within the European Union. Eur. J. Pharm. Biopharm. 2018, 131, 8–24.
[CrossRef] [PubMed]
4. Arakawa, T.; Prestrelski, S.J.; Kenney, W.C.; Carpenter, J.F. Factors affecting short-term and long term
stabilities of proteins. Adv. Drug Deliv. Rev. 1993, 10, 1–28. [CrossRef]
5. Franks, F. Long-term stabilization of biologicals. Biotechnology 1994, 12, 253–256. [CrossRef]
6. Manning, M.C.; Patel, K.; Borchardt, R.T. Stability of Protein Pharmaceuticals. Pharm. Res. 1989, 6, 903–918.
[CrossRef]
7. Manning, M.C.; Chou, D.K.; Murphy, B.M.; Payne, R.W.; Katayama, D.S. Stability of protein pharmaceuticals:
An update. Pharm. Res. 2010, 27, 544–575. [CrossRef]
8. Tang, X.; Pikal, M.J. Design of Freeze-Drying Processes for Pharmaceuticals: Practical Advice. Pharm. Res.
2004, 21, 191–200. [CrossRef]
9. Patel, S.M.; Pikal, M.J. Emerging freeze-drying process development and scale-up issues. AAPS Pharmscitech
2011, 12, 372–378. [CrossRef]
10. Bjelošević, M.; Seljak, K.B.; Trstenjak, U.; Logar, M.; Brus, B.; Ahlin Grabnar, P. Aggressive conditions during
primary drying as a contemporary approach to optimise freeze-drying cycles of biopharmaceuticals. Eur. J.
Pharm. Sci. 2018, 122, 292–302. [CrossRef]
11. Depaz, R.A.; Pansare, S.; Patel, S.M. Freeze-Drying Above the Glass Transition Temperature in Amorphous
Protein Formulations While Maintaining Product Quality and Improving Process Efficiency. J. Pharm. Sci.
2015, 105, 40–49. [CrossRef] [PubMed]
12. Gitter, J.H.; Geidobler, R.; Presser, I.; Winter, G. A Comparison of Controlled Ice Nucleation Techniques for
Freeze-Drying of a Therapeutic Antibody. J. Pharm. Sci. 2018, 107, 2748–2754. [CrossRef] [PubMed]
13. Zhang, M.; Jiang, H.; Lim, R.X. Recent developments in microwave-assisted drying of vegetables, fruits, and
aquatic products-drying kinetics and quality considerations. Dry. Technol. 2010, 28, 1307–1316. [CrossRef]
14. Jiang, H.; Zhang, M.; Mujumdar, A.S. Physico-chemical changes during different stages of MFD/FD banana
chips. J. Food Eng. 2010, 101, 140–145. [CrossRef]
15. Meredith, R. Engineers’ Handbook of Industrial Microwave Heating; Reprint wi; The Institution of Engineering
and Technology: London, UK, 2007; ISBN 978-0-85296-916-8.
16. Metaxas, A.C. Microwave heating. Power Eng. J. 1991, 5, 237–247. [CrossRef]
17. Datta, A.K.; Davidson, P.M. Microwave and Radio Frequency Processing. J. Food Sci. 2000, 65, 32–41.
[CrossRef]
18. Chandrasekaran, S.; Ramanathan, S.; Basak, T. Microwave food processing—A review. Food Res. Int. 2013,
52, 243–261. [CrossRef]
19. Ambros, S.; Mayer, R.; Schumann, B.; Kulozik, U. Microwave-freeze drying of lactic acid bacteria: Influence
of process parameters on drying behavior and viability. Innov. Food Sci. Emerg. Technol. 2018, 48, 90–98.
[CrossRef]
20. Gitter, J.H.; Geidobler, R.; Presser, I.; Winter, G. Significant Drying Time Reduction Using Microwave-Assisted
Freeze-Drying for a Monoclonal Antibody. J. Pharm. Sci. 2018, 107, 2538–2543. [CrossRef]
21. Duan, X.; Zhang, M.; Mujumdar, A.S.; Wang, S. Microwave freeze drying of sea cucumber (Stichopus
japonicus). J. Food Eng. 2010, 96, 491–497. [CrossRef]
22. Jiang, H.; Zhang, M.; Liu, Y.; Mujumdar, A.S.; Liu, H. The energy consumption and color analysis of
freeze/microwave freeze banana chips. Food Bioprod. Process. Trans. Inst. Chem. Eng. Part C 2013, 91, 464–472.
[CrossRef]
23. Duan, X.; Zhang, M.; Mujumdar, A.S. Studies on the Microwave Freeze Drying Technique and Sterilization
Characteristics of Cabbage. Dry. Technol. 2007, 25, 1725–1731. [CrossRef]
24. Evans, R. Applications of Microwave Vacuum Drying and Lyospheres to Freeze-Drying of Vaccines and
Biologics. In Proceedings of the CPPR: Freeze-Drying of Pharmaceuticals and Biologicals Conference,
Garmisch-Partenkrichen, Germany, 23–26 September 2014.
25. Durance, T.D.; Fu, J. Apparatus and Method for Dehydrating Biological Materials. World Intellectual
Property Organization Patent Application No. WO2010124375A1, 4 November 2010.
26. Bhambhani, A.; Evans, R.K.; Sinacola, J.; Jones, M. Method of Obtaining Thermostable Dried Vaccine
Formulations. World Intellectual Property Organization Patent Application No. WO 2015/057541 A1, 23
April 2015.
27. Bhambhani, A.; Evans, R.K.; Sinacola, J.; Lizzano, R. Method of Microwave Vacuum Drying Spherical-shaped
Pellets of Biological Materials. World Intellectual Property Organization Patent Application No.
WO2015057540A1, 23 April 2015.
28. Püschner GmbH & Co. KG. Good Manufacturing Practise (GMP) for Industrial Microwave Plants. Available
online: http://www.pueschner.com/en/microwave-technology/gmp-microwaves (accessed on 10 April 2018).
29. Sundaramurthi, P.; Suryanarayanan, R. Trehalose crystallization during freeze-drying: Implications on
lyoprotection. J. Phys. Chem. Lett. 2010, 1, 510–514. [CrossRef]
30. Sundaramurthi, P.; Suryanarayanan, R. Influence of crystallizing and non-crystallizing cosolutes on trehalose
crystallization during freeze-drying. Pharm. Res. 2010, 27, 2384–2393. [CrossRef]
31. Liu, J. Physical characterization of pharmaceutical formulations in frozen and freeze-dried solid states:
Techniques and applications in freeze-drying development. Pharm. Dev. Technol. 2006, 11, 3–28. [CrossRef]
32. Searles, J.A.; Carpenter, J.F.; Randolph, T.W. The ice nucleation temperature determines the primary drying
rate of lyophilization for samples frozen on a temperature-controlled shelf. J. Pharm. Sci. 2001, 90, 860–871.
[CrossRef]
33. Vadivambal, R.; Jayas, D.S. Non-uniform temperature distribution during microwave heating of food
materials—A review. Food Bioprocess Technol. 2010, 3, 161–171. [CrossRef]
34. Burfoot, D.; Griffin, W.J.; James, S.J. Microwave pasteurisation of prepared meals. J. Food Eng. 1988, 8,
35. Vilayannur, R.S.; Puri, V.M.; Anantheswaran, R.C. Size and shape effect on nonuniformity of temperature
and moisture distributions in microwave heated food materials: Part II experimental validation. J. Food
Process Eng. 1998, 21, 235–248. [CrossRef]
36. Zhang, Z.; Su, T.; Zhang, S. Shape Effect on the Temperature Field during Microwave Heating Process. J.
Food Qual. 2018, 2018, 1–24. [CrossRef]
37. James, C.; Swain, M.V.; James, S.J.; Swain, M.J. Development of methodology for assessing the heating
performance of domestic microwave ovens. Int. J. Food Sci. Technol. 2002, 37, 879–892. [CrossRef]
38. Rattanadecho, P. Theoretical and experimental investigation of microwave thawing of frozen layer using a
microwave oven (effects of layered configurations and layer thickness). Int. J. Heat Mass Transf. 2004, 47,
39. Lobo, S.; Datta, A.K. Characterization of Spatial Non-Uniformity in Microwave Reheating of High Loss
Foods. J. Microw. Power Electromagn. Energy 1998, 33, 158–166. [CrossRef]
40. Pikal, M.J.; Shah, S. Moisture transfer from stopper to product and resulting stability implications. Dev. Biol.
Stand. 1992, 74, 165–177. [PubMed]
41. Meng, C.; Da-ming, F.; Lue-lue, H.; Yi-shu, G.; Jian-lian, H.; Jian-xin, Z.; Hao, Z. A new approach to microwave
food research: Analyzing the electromagnetic response of basic amino acids. Innov. Food Sci. Emerg. Technol.
2017, 41, 100–108. [CrossRef]
42. Aoki, K.; Nagano, K.; Iitaka, Y. The crystal structure of L-arginine phosphate monohydrate. Acta Crystallogr.
Sect. B Struct. Crystallogr. Cryst. Chem. 1971, 27, 11–23. [CrossRef]
43. Izutsu, K.I.; Fujimaki, Y.; Kuwabara, A.; Aoyagi, N. Effect of counterions on the physical properties of
L-arginine in frozen solutions and freeze-dried solids. Int. J. Pharm. 2005, 301, 161–169. [CrossRef]
44. Stärtzel, P.; Gieseler, H.; Gieseler, M.; Abdul-Fattah, A.M.; Adler, M.; Mahler, H.-C.; Goldbach, P. Freeze
Drying of L-Arginine/Sucrose-Based Protein Formulations, Part I: Influence of Formulation and Arginine
Counter Ion on the Critical Formulation Temperature, Product Performance and Protein Stability. J. Pharm.
Sci. 2015, 104, 2345–2358. [CrossRef]
45. Pikal, M.J.; Shah, S.; Roy, M.L.; Putman, R. The secondary drying stage of freeze drying: Drying kinetics as a
function of temperature and chamber pressure. Int. J. Pharm. 1990, 60, 203–217. [CrossRef]
46. Bianchi, C.; Schmid, R.; Frick, D.; Kurtz, S. Microwave Assisted Vacuum Drying Processing: Magnetron vs
Solid State. Case Study: Apple Drying. In Proceedings of the COMSOL Conference, Lausanne, Switzerland,
22–24 October 2018.
47. Wang, W. Lyophilization and development of solid protein pharmaceuticals. Int. J. Pharm. 2000, 203, 1–60.
[CrossRef]
48. Du, Y.; Walsh, A.; Ehrick, R.; Xu, W.; May, K.; Liu, H. Chromatographic analysis of the acidic and basic
species of recombinant monoclonal antibodies. MAbs 2012, 4, 578–585. [CrossRef] [PubMed]
49. Fekete, S.; Beck, A.; Veuthey, J.L.; Guillarme, D. Ion-exchange chromatography for the characterization of
biopharmaceuticals. J. Pharm. Biomed. Anal. 2015, 113, 43–55. [CrossRef] [PubMed]
50. Harris, R.J.; Kabakoff, B.; Macchi, F.D.; Shen, F.J.; Kwong, M.; Andya, J.D.; Shire, S.J.; Bjork, N.; Totpal, K.;
Chen, A.B. Identification of multiple sources of charge heterogeneity in a recombinant antibody. J. Chromatogr.
B Biomed. Sci. Appl. 2001, 752, 233–245. [CrossRef]
51. Stärtzel, P. Arginine as an Excipient for Protein Freeze-Drying: A Mini Review. J. Pharm. Sci. 2018, 107,
960–967. [CrossRef] [PubMed]
52. Abdul-Fattah, A.M.; Truong-Le, V.; Yee, L.; Nguyen, L.; Kalonia, D.S.; Cicerone, M.T.; Pikal, M.J.
Drying-Induced Variations in Physico-Chemical Properties of Amorphous Pharmaceuticals and Their
Impact on Stability (I): Stability of a Monoclonal Antibody. J. Pharm. Sci. 2007, 96, 1983–2008. [PubMed]
53. Craig, D.Q.M.; Royall, P.G.; Kett, V.L.; Hopton, M.L. The relevance of the amorphous state to pharmaceutical
dosage forms: Glassy drugs and freeze dried systems. Int. J. Pharm. 1999, 179, 179–207. [CrossRef]
54. Yoshioka, S.; Aso, Y. Correlations between Molecular Mobility and Chemical Stability during Storage of
Amorphous Pharmaceuticals. J. Pharm. Sci. 2007, 96, 960–981. [CrossRef]
55. Abdul-Fattah, A.M.; Kalonia, D.S.; Pikal, M.J. The Challenge of Drying Method Selection for Protein
Pharmaceuticals: Product Quality Implications. J. Pharm. Sci. 2007, 96, 1886–1916. [CrossRef]
56. Schersch, K.; Betz, O.; Garidel, P.; Muehlau, S.; Bassarab, S.; Winter, G. Systematic Investigation of the Effect
of Lyophilizate Collapse on Pharmaceutically Relevant Proteins, Part 2: Stability During Storage at Elevated
Temperatures. J. Pharm. Sci. 2012, 101, 2288–2306. [CrossRef]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Microwave-Assisted Freeze-Drying of Monoclonal Antibodies: Product Quality Aspects and Storage Stability

Uploaded by

Copyright:

Available Formats

Microwave-Assisted Freeze-Drying of Monoclonal Antibodies: Product Quality Aspects and Storage Stability

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Microwave-Assisted Freeze-Drying of Monoclonal Antibodies: Product Quality Aspects and Storage Stability

Uploaded by

Copyright:

Available Formats

pharmaceutics

Keywords: freeze-drying; lyophilization; drying; microwave; protein; monoclonal antibody; stability

Pharmaceutics 2019, 11, 674; doi:10.3390/pharmaceutics11120674 www.mdpi.com/journal/pharmaceutics

2. Materials and Methods

2.2. Study Design

Table 1. Formulations used in this study.

2.3. Preparation of Formulations

2.4. Freeze-Drying Process

2.4.1. Conventional Freeze-Drying (CFD)

Freeze-Drying Process Setpoint Freezing Primary Drying Secondary Drying

2.4.2. Microwave-Assisted Freeze-Drying (MFD)

In contrast, F4 was processed with a partially modified setup comprising a semiconductor

2.6. Brunauer-Emmet-Teller Krypton Gas Adsorption

2.6. Brunauer-Emmet-Teller Krypton Gas Adsorption

2.7. X-Ray Powder Diffraction

2.8. Reconstitution of Lyophilizates

2.9. High-Performance Size Exclusion Chromatography (HP-SEC)

2.10. High-Performance Cation Exchange Chromatography (HP-CEX)

2.11. Light Obscuration

2.12. Flow-Imaging Microscopy

3.1. Solid State

3.1.1. Residual Moisture Content and Specific Surface Area

3.1.2. X-Ray Powder Diffraction (XRD)

3.1.2. X-Ray Powder Diffraction (XRD)

3.2. Protein-Related Quality Attributes

3.2.1. Reconstitution and Subvisible Particles (SvP)

Table 3. Reconstitution times of the different formulations.

Formulation Reconstitution Time (s)

3.2.2. Weak Cation Exchange Chromatography (CEX)

3.2.2. Weak Cation Exchange Chromatography (CEX)

4.1. Stability with Regard to Solid State Properties

4.2. Stability with Regard to Protein-Related Properties

Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4923/11/12/674/s1,

You might also like